Meta-Auxiliary Learning for Future Depth Prediction in Videos

Liu, Huan; Chi, Zhixiang; Yu, Yuanhao; Wang, Yan; Tang, Jin

doi:10.1109/wacv56688.2023.00571

Cited by 9 publications

(1 citation statement)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our use of meta-learning is different than that in (Park et al 2019). Instead, our meta-learning approach is closer to a variant of MAML, i.e., meta-auxiliary learning (Liu, Davison, and Johns 2019;Chi et al 2021;Liu et al 2022Liu et al , 2023a. It aims to address the primary task by solving an auxiliary task.…”

Section: Meta-learningmentioning

confidence: 99%

Test-Time Personalization with Meta Prompt for Gaze Estimation

Liu,

Qi,

et al. 2024

AAAI

View full text Add to dashboard Cite

Despite the recent remarkable achievement in gaze estimation, efficient and accurate personalization of gaze estimation without labels is a practical problem but rarely touched on in the literature. To achieve efficient personalization, we take inspiration from the recent advances in Natural Language Processing (NLP) by updating a negligible number of parameters, "prompts", at the test time. Specifically, the prompt is additionally attached without perturbing original network and can contain less than 1% of a ResNet-18's parameters. Our experiments show high efficiency of the prompt tuning approach. The proposed one can be 10 times faster in terms of adaptation speed than the methods compared. However, it is non-trivial to update the prompt for personalized gaze estimation without labels. At the test time, it is essential to ensure that the minimizing of particular unsupervised loss leads to the goals of minimizing gaze estimation error. To address this difficulty, we propose to meta-learn the prompt to ensure that its updates align with the goal. Our experiments show that the meta-learned prompt can be effectively adapted even with a simple symmetry loss. In addition, we experiment on four cross-dataset validations to show the remarkable advantages of the proposed method.

show abstract